What Are The Bluesky Bots Doing? — BCB #130
Also: building bridges using existing volunteer networks, and trust in elections increases, sort of.
In the wake of the election, liberals in particular have been migrating en masse away from Elon Musk’s X and over to Bluesky. (A million new accounts were created in the first week after the election alone.) The new platform purports to be a home for healthier discourse. But some users are finding that swarms of argumentative, hostile bots are popping up in their Bluesky replies, too. Bluesky user DuendeonFuego documents this phenomenon in a recent thread:
Earlier this month, longtime YouTuber Hank Green vlogged about the issue, showing viewers how he’d encountered a bot account on Bluesky intent on disagreeing with as many people as possible, in response to hot-button political takes and also seemingly innocuous ones. Green asked his viewers to help him understand what the point of these accounts is:
If there’s somebody out here on the internet making me mad, there’s a high probability that that’s not even a real person… What’s the incentive? The only incentive is to take up my time, distract me, and make me have less faith in my society and fellow people.
This sounds like a strange thing to do, but there are actual incentives to do it. While it is well known that Russian bots tried to help Trump win in 2016, they also had a strategy of just increasing polarization generally by posting inflammatory content on both sides of issues like gun control and race relations. Reducing social cohesion can itself be a political goal, because polarization serves conflict entrepreneurs by increasing engagement.
Or perhaps there’s some person or group who wants to attack Bluesky as a platform. It’s easy to imagine the abuse of “inclusive” language as someone’s spiteful project, though we don’t want to jump to conclusions. DuendeonFuego writes that these argumentative and hateful accounts can serve as “vibe pollution,” over time spurring users to lock their accounts or leave the site entirely:
The seemingly random bot networks are an intentional effort to toxify online spaces and keep people angry and fighting. If these can create enough ambient hostility, they can degrade the utility of online platforms and/or destroy them entirely.
But we don’t really know the extent of the issue, or what’s actually going on here. There are at least two other plausible explanations.
These bot accounts may simply be “engagement farming,” trying to build up followers and long-term reputation so that they can eventually be sold to marketers or propagandists. We’ve repeatedly seen this on other platforms, where there’s a gray market for aged accounts—the age of an account is one important credibility signal used by humans and machines alike to try to determine whether a source can be trusted.
Alternatively, this could be some sort of research. These bots could be intentionally disagreeable just to see how people respond. On the other hand, the tone is ambiguous—are they passive-aggressive or sincere in suggesting inclusiveness? There’s a chance these bots could be part of an experiment that is genuinely trying to help people that is somehow going wrong. Previous research into using AI-generated replies to curb hate speech has found that this method can backfire, increasing hostility.
It’s early days still for Bluesky. Maybe its architects will figure out how to quash bots without killing broader conversation. But as Bluesky continues to increase in popularity, it will only become a more popular target. A big part of what makes a platform successful is what you don’t see—the security measures that keep spammers and bots out. In the age of AI, this will only get more difficult.
A proposal for uniting America at scale
In a recent blog post, conflict resolution guru Peter T. Coleman makes a compelling case for using the infrastructure of existing volunteer service programs to take steps towards reuniting the country.
Coleman builds on conversations with leaders from organizations like Habitat for Humanity, the Red Cross, YMCA, and Americorps, and research showing that a common purpose can dissolve perceived differences remarkably effectively. Given that more than 60 million Americans volunteer regularly through these and other organizations, using these networks would be an amazing way to quickly involve large numbers of people in bridge building work.
For this to work though, programs must be intentionally designed and thought-out. As Coleman explains:
Organizations can’t just throw people together and hope for the best. The most successful programs ensure participants have equal status, sustained contact over time, and genuine interdependence in achieving their goals. They’re intentional in their pairing of volunteers, and thoughtful about local community dynamics and power relationships. Most importantly, they integrate bridge-building organically into existing service work rather than creating artificial “kumbaya” moments. People don’t have to sign up explicitly for bridge-building — they can simply show up to serve their community and find themselves working productively with people they might otherwise never meet.
There’s a lot to be said for working with what we have rather than trying to create something totally new. As Coleman adds, there’s already momentum building behind this approach:
Last April, bipartisan legislation was proposed in Congress that would provide $25 million annually to AmeriCorps specifically for integrating bridge-building into service programs. A new Trust for Civic Infrastructure also launched with $30 million in initial funding to support community bridge-building work. Last August, a report was released, No Greater Mission. No Greater Means. How National Service Can Advance Bridgebuilding, which shared grounded insights from this initiative that national service organizations are currently exploring.
Trust in elections is improving—with an asterisk
Whether or not people were happy with the outcome of this year’s election, few are disputing the legitimacy of Trump’s victory. Post-election polling has found that only 10 percent of the losing side did not accept the election results as legitimate. This is a marked shift from 2020, when 70 percent of Republicans questioned the legitimacy of Biden’s victory, and from 2016, where 56 percent of Democrats did not accept Trump as the legitimate president.
This year, new research from Pew finds that the overwhelming majority of Americans across the board felt confident that elections were well run, and that both in-person and mail-in ballots were counted accurately. As political commentator Richard Hanania put it, “Republican election denial made Democrats accept all elections. A rare case of affective polarization making a side better.”
There was, however, a slight decrease in Blue trust in election results. This is consistent with historic patterns, where the losing side tends to have lower trust in elections.
Quote of the Week
Posting a statement only on BlueSky (Cheney, AOC) is now its own statement.